Occurrence, Dissipation Kinetics and Environmental Risk Assessment of Antibiotics and their Metabolites in Agricultural Soils

Journal of Hazardous Materials 479 (2024) 135586 Available online 22 August 2024 0304-3894/© 2024 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Occurrence, dissipation kinetics and environmental risk assessment of antibiotics and their metabolites in agricultural soils Carmen Mejías, Laura Martín-Pozo, Juan Luis Santos * , Julia Martín , Irene Aparicio , Esteban Alonso Departamento de Química Analítica, Escuela Polit´ ecnica Superior, Universidad de Sevilla, C/ Virgen de ´ Africa, 7, E-41011 Seville, Spain HIGHLIGHTS GRAPHICAL ABSTRACT •Experimental data revealed biphasic kinetics for most compounds •Metabolites dissipated more quickly than their parent compounds •Trimethoprim was more persistent in the soils than sulfonamides •DT 50 values were from 0.09 to >120 days •Environmental risk assessment suggested no risk to terrestrial organisms Graphical abstract ARTICLE INFO Keywords: Antibiotics Metabolites Agricultural soils Occurrence Degradation ABSTRACT Antibiotics are among the emerging contaminants of greatest concern to the scientific community. However, the occurrence and behaviour of their metabolites in soils have been scarcely studied. To address this research gap, this study investigates the occurrence, sorption, dissipation kinetics, and environmental risk of highly important antibiotics (sulfamethazine, sulfadiazine, sulfamethoxazole, trimethoprim) and their main metabolites in Mediterranean agricultural soils. Batch experiments were conducted under natural conditions for 120 days. Five different dissipation kinetics models were applied to elucidate antibiotics degradation. The sorption isotherms were evaluated by three different models. Most of the antibiotics and metabolites tested showed a good fit with the Linear Isotherm model (R 2 >0.96) and biphasic dissipation kinetic models (R 2 >0.90). The dissipation and the endpoints values (DT 50 and DT 90 ) depended on the soil type properties. A Lixisol soil demonstrated reduced degradation of the investigated compounds. Trimethoprim showed the highest persistence, followed by sulfamethazine, sulfamethoxazole, and sulfadiazine. Parent compounds exhibited lower degradation rates than their metabolites. Remaining antibiotic concentrations were found to be below the predicted no-effect concentration in soil, suggesting that they may not pose a risk to terrestrial biota. This study provides valuable insights into the behaviour of these antibiotics and their metabolites in soil. * Corresponding author. E-mail address: [email protected] (J.L. Santos). Contents lists available at ScienceDirect Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat https://doi.org/10.1016/j.jhazmat.2024.135586 Received 21 May 2024; Received in revised form 8 August 2024; Accepted 19 August 2024 Journal of Hazardous Materials 479 (2024) 135586 2 1. Introduction Pharmaceutically active compounds (PhACs) are the emerging contaminants which have received greater attention in recent years, due to their potential biological activity and their continuous consumption and discharge to the environment [1,2]. After this consumption, PhACs reach to the wastewater treatment plants (WWTPs) through sewer system. WWTPs are not designed to remove these pollutants and most of them remain in effluent wastewater or/and are retained in sewage sludge generated during the wastewater treatment. The use of treated wastewater, or even receiving waters, for irrigation, and the disposal of sewage sludge or manure as fertilizer in agricultural lands, are the main disposal routes of PhACs in the soil [3,4]. Their occurrence in soil environment is mainly governed by sorption/desorption processes [5-7], and transformation or microbiological degradation [4,8]. PhACs could affect soils microorganisms [2] or crop yields [1] and even they could be taken by plants, entering the food chain by the contamination of crops [9]. Antibiotics comprise a group of PhACs utilized in both human and veterinary medicine. This group includes fluoroquinolones, macrolides, tetracyclines or sulfonamides among other, which are widely used in the treatment of bacterial infections and as growth promoters. The main concern associated with antibiotics is their ability to promote the creation of antibiotic resistance genes and antibiotic resistant bacteria [10]. These compounds have showed up in agriculture soils at concentrations even up to milligram per kilogram [11,12,3]. At these concentrations, antibiotics could negatively affect soil microbiome and crops. For example, Bao et al. [13] reported that environmental concentrations of erythromycin could lead to growth retardation and antioxidant defense of pakchoi. Moreover, the accumulation of antibiotic in plants would lead a risk to public health due to direct human exposure through the food chain [14]. The recent literature about the occurrence of antibiotics in soil has been mainly focused on the evaluation of leaching, lixiviation, sorption, and desorption processes [15], while only a few studies have evaluated the dissipation of these compounds in soil. Moreover, most of these studies use a single-first order kinetic model (SFO) to evaluate the dissipation of these compounds in soil [16-19]. However, the dissipation of the organic pollutants is strongly influenced by their availability in soil which depend on the physicochemical characteristics of the compound (as pK a or Log K ow ) and the soil (texture, organic matter content (OM), pH, etc.). For example, Kulesza et al. [20] observed a faster dissipation of the antibiotic pirlimycin in fine-textured soil with high OM than those observed in coarse-textured soil with low OM after the application of cow manure to the soil. Other studies have showed the influence of these parameters in the degradation of antibiotics as tetracycline [21] or sulfonamides [22]. As result, the use of SFO model to evaluate the dissipation kinetic for all compounds and soils could result on an under- or overestimation of the dissipation endpoints (DT 50 and DT 90 ). Actually, previous studies have showed a better fit of the dissipation kinetic of antibiotics using biphasic models than those using the SFO kinetic model [4,23]. Consequently, significant differences were obtained in the endpoints obtained between models [4]. In addition, the occurrence of antibiotic metabolites has been poorly explored, which represent a critical knowledge gap [24-27,4]. This is particularly significant, as metabolites could be more toxic and show higher concentrations than their parent compounds [26]. As a consequence, there is lack of information regarding the occurrence, sorption and especially the dissipation of these compounds in environmental soils. Moreover, evaluating these metabolites is crucial for understanding the biotransformation processes of parent compounds into metabolites by soil microorganisms, or vice versa. Taking into account these premises, the aim of this work was to determinate the dissipation of the most representative sulfonamide antibiotics, trimethoprim (TMP) and their main metabolites in Mediterranean soils with different physical-chemical properties under natural conditions, evaluate the kinetics of the process, and assess their potential environmental risks. For this purpose, three sulfonamides, TMP and six of their main metabolites were selected as the target compounds. Selected antibiotics were classified as highly important classes of antibiotics by World Health Organization. Metabolites, to date, have been poorly studied in soils, which may pose more risk than the parent compounds. To the best of our knowledge, this is the first time that data on the dissipation kinetics and half-life times of sulfonamide and trimethoprim metabolites have been reported. Metabolites were selected according to their ratio of formation to the parent compound, thus selecting the metabolites that are formed in higher extent after metabolism. Sulfonamides human metabolism is performed mainly by arylamine N-acetyltransferase enzymes, which are responsible for its acetylation at its N 4 position. Glucuronidation of the N 4 atom is an additional route of metabolism of sulfamethoxazole (SMX) [28]. After intake, 80% of TMP is still unaltered, while the other 20% is metabolized. These metabolites are produced mainly by demethylation (90%) of the parent compound, by its hydroxylation and oxidation [29]. 2. Experimental 2.1. Chemicals and reagents Standards of sulfamethazine (SMZ, ≥99.0%), N 4 -acetylsulfame- thoxazole (AcSMX, ≥98.5%), N 4 -acetylsulfadiazine (AcSDZ, >99.0%) and sulfadiazine (SDZ, ≥99.0%) were obtained from Sigma-Aldrich (Steinheim, Germany). N 4 -acetylsulfamethazine (AcSMZ, ≥98.0%), sulfamethoxazole N 4 -glucoside (SMX-GL, ≥99.0%), 3-desmethyltrime- thoprim (DM-TMP, ≥98.0%) and 4-hydroxytrimethoprim (4-OH-TMP, ≥97.0%) were provided from Toronto Research Chemicals (Toronto, Canada). SMX (≥99.5%) and TMP (≥99.5%) were supplied by Dr Ehrenstorfer GmbH (Augsburg, Germany). Erythromycin-(N,N-dime- thyl- 13 C 2 ) (ERY- 13 C, ≥99.0%) and sulfamethoxazole-(phenyl- 13 C 6 ) (SMX- 13 C, ≥99.0%) were supplied by Sigma-Aldrich (Steinheim, Germany). The physicochemical characteristics of target antibiotics are shown in Table 1. Ammonium formate was provided from Sigma- Aldrich (Steinheim, Germany). C18 sorbent was supplied by Scharlab (Barcelona, Spain). Formic acid and analytical-grade calcium chloride anhydrous (CaCl 2 ) were purchased from Panreac (Barcelona, Spain). Methanol and water (LC-MS-grade) were supplied by Merck (Darmstadt, Germany). Individual standard solutions were prepared in methanol at 1000 mg L −1 . Mixed standard solutions were prepared by dilution. The XLSTAT add-in for Excel was used for the Partial Least Squares Regression (PLS) analysis. 2.2. Sample collection Considering the potential human risks due to the translocation of studied compounds into the crops, the three most widely distributed Mediterranean agricultural soils were collected from Seville (South of Spain). The soils showed different physical and chemical characteristics. According to the World Reference Base for Soil Resources [35], soil 1 was a Mollic Fluvisol, an Alluvial nature soil sited in the plain of the main rivers from Europe. This soil usually shows high content of silt and clay. Soil 2 was a Rhodic Lixisol, a red soil with high drainage capacity and, soil 3 was a Cambic Cambisol soil. Soils 2 and 3 are widely extended in several countries from Europe such as Spain, Italy, Greek, France, or Germany. For each soil, the surface soils (0–20 cm) were sampled, freeze-dried in a Cryodos-50 lyophilizer (Telstar, Terrasa, Spain), homogenized, sieved (particle size <2 mm), and stored in glass bottles at −18 ◦C for one week until the batch experiments. The three soils were characterized. The applied procedures are in supplementary materials and results are shown in Table 2. Soil 1 showed high silt and clay while soils 2 and 3 showed high sand content. According to their texture classes, soil 1 was classified as Clay loam (CL), soil 2 as Loamy sand (LS), and soil 3 as Sandy loam (SL). C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 3 2.3. Sorption isotherm experiments Sorption assays were conducted in triplicate following the guidelines outlined in OECD guideline 106 [36]. Falcon 50 mL centrifuge tubes served as the experimental vessels. All assays were carried out at a temperature of 25ºC. To assess the potential degradation, as well as any sorption onto the plastic of the falcon tubes or onto the nylon filter, quality control experiments were conducted without soil. The soil-solution mixture was agitated using a rotator shaker (LLG-uniLOOPMIX2) set at 40 rpm. The sorption isotherms were assessed across the three designated soil types. To initiate this, 5 g samples of each soil variant were placed into separate falcon centrifuge tubes, followed by the addition of 9.5 mL of 0.01 M CaCl 2 aqueous solution. This soil-solution mixture underwent agitation for 24 h to achieve pre-equilibration of the soil. Subsequently, 0.5 mL of various standard solutions, prepared in CaCl 2 aqueous solution, were introduced, yielding final concentrations of 0.01, 0.025, 0.05, 0.1, 0.25, and 0.5 mg L −1 (for sulfonamides and their metabolites) and 0.8, 1.2, 2.0, Table 1 Characteristics of target compounds. Compound CAS number Molecular weight (g mol ¡1 ) pK a Water solubility (mg mL ¡1 ) Log K ow Structure Trimethoprim (TMP) 738 −70 −5 290.3 7.16 a (strongest acidic), 17.33 a (strongest basic) 0.615 a 0.91 g 4-hydroxytrimetoprim (4-OH- TMP) 112678 −48 −5 306.3 8.18 b n.a. n.a. 3-desmethyltrimethoprim (DMTMP) 27653 −69 −6 276.3 9.40 b n.a. n.a. Sulfamethoxazole (SMX) 723 −46 −6 253.3 1.97 a (strongest basic), 6.16 a (strongest acidic) 0.459 a 0.89 g N 4 -acetylsulfamethoxazole (AcSMX) 21312 −10 −7 295.3 5.54 e 0.076 b 1.18 e Sulfamethoxazole N 4 -glucoside (SMX-GL) 119691 −75 −7 415.4 n.a. n.a. n.a. Sulfadiazine (SDZ) 68 −35 −9 250.3 2.01 a (strongest basic), 6.99 a (strongest acidic) 0.601 −0.09 g N 4 -acetylsulfadiazine (AcSDZ) 127 −74 −2 292.3 6.1 d 0.15 b 0.39 f Sulfamethazine (SMZ) 57 −68 −1 278.3 2.04 a (strongest basic), 6.99 a (strongest acidic) 0.23 a 0.25 g N 4 -acetylsulfamethazine (AcSMZ) 100 −90 −3 320.4 7.16 c 0.929 b 1.48 f Parent compounds are marked in bold; Abbreviations are written in brackets; n.a.: not available. a : [16]; b : [14]; c : [30]; d : [31]; e : [32]; f : [33]; g : [34]. C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 4 3.2, 4.8, and 8.0 mg L −1 (for TMP and its metabolites) in a total volume of 10 mL. The centrifuge tubes were then agitated for an additional 24 h. After the sorption experiment, the tubes were centrifuged at 5000 rpm for 10 min. The liquid phase was filtered through a 0.22 µm nylon filter. Analysis of antibiotics was performed by direct injection with liquid-chromatography tandem mass spectrometry (LC-MS/MS). 2.4. Batch dissipation experiments Batch experiments were carried out in the three selected soils according to OECD guideline 307 [37]. All the studied soils were previously analyzed to ensure the absence of the target compounds. The experiments were done in non-contaminated soils spiked with target compounds (antibiotics and their metabolites) at a concentration of 500 μ g kg −1 dry matter (dm) using an aqueous solution and under climatic conditions controlled by a climatic chamber. This concentration was selected based on previous dissipation studies [4,26] and is consistent with those contents of antibiotics and metabolites typically found in the environment, usually measured in soils, sludge and wastewaters applied to the soils [38-42]. Therefore, 100 mL of working solution containing 2 mg L −1 of the studied compounds was added to 400 g of each soil. The mixtures were homogenized for 2 h in a rotator shaker (LLG-uniLOOPMIX2) at 40 rpm. For each soil, batch experiments were carried out by quintuplicate (5 glass reactors), and two glass reactors were fixed with non-spiked soil as blanks (100 mL of distilled water was added to 400 g of soil). Experiments were carried out under aerobic conditions in a Hotcold GL climate chamber (Selecta, Barcelona, Spain) at 25ºC with 33 % of relative humidity, 20 % of soil humidity and irradiation levels of 1091 W/m 2 in all experiments. These conditions were selected according to the typical Mediterranean climate. Soil homogeneity was directly measured through the variation in soil resistivity by means of a multiparametric probe that was buried in the soil (Hanna Instruments) before the spiking procedure. Soil humidity was checked every sampling day and corrected by the addition of distilled water. Forty samples (5.0 g) were collected from each reactor as follows: five weekly samples were collected during the first fourth weeks (from days 0–27); three weekly samples from days 28–62, a weekly sample from days 63–90; two samples at the following 15 days (after 105 and 120 days). Samples were collected from different points of the reactor and mixed, always after homogenization of the reactor. 2.5. Analytical determination Collected soils samples were freeze-dried in a Cryodos-50 lyophilizer (Telstar, Terrasa, Spain). After lyophilization, samples were homogenized, crushed in a mortar, and sieved (<2 mm). The analysis of target compounds was carried out according to Mejías et al. [43]. Briefly, sample extraction (1 g) was carried out by ultrasound-assisted extraction with 3 mL of methanol 0.5 % formic acid (v/v), the extract clean-up by dispersive solid-phase extraction with 0.4 g of C18, and the determination by liquid-chromatography tandem mass spectrometry, according to a previous published methodology [43]. Chromatographic analyses were performed using an Agilent 1290 Infinity II chromatograph (Agilent, USA). Chromatographic conditions are widely described in the supplementary material. Table S1 shows LC-MS/MS parameters of target compounds. In Table S2 can be seen the validation parameters obtained from the method. 2.6. Data analysis 2.6.1. Dissipation kinetic FOCUS guidance document on estimating persistence and degradation kinetics from environmental fate studies on pesticides [44] was employed for the assessment of the dissipation kinetics. Dissipation kinetics and the estimation of the endpoints for antibiotics and their metabolites (DT 50 and DT 90 , corresponding to the dissipation of the 50 % and 90 % of the initial concentrations, respectively) was evaluated by the following kinetic models: a SFO model; three biphasic models: bi-exponential model (BEM), first-order two-compartment (FOTC) model, and first-order double-exponential decay (FODED) model; and a model suitable to describe the dissipation kinetics with a lag phase: a logistic model (LM) [4]. Models and their equations are widely described in Table 3. Parameters of each of the studied models were optimized using Microsoft Excel-Solver, by minimizing the mean of the square error, that is, the sum of the square of the residuals (SSR) divided by the number of experimental data. The fit of each model to the experimental data was first evaluated through visual assessment by representing the experimental mean concentrations vs the concentrations calculated for each model. The goodness fit of the model was assessed considering the coefficient of determination (R 2 ) and the curve slope (close to 1) of the linear curve. DT 50 and DT 90 values were calculated for each kinetic model. In the SFO model, DT 50 and DT 90 were calculated using Eqs. (1) and (2) as follows: DT 50 =ln 2 / k 1 (1) DT 90 =ln 10 / k 1 (2) For the other models, there were no specific equations available for estimating DT 50 and DT 90 . Therefore, these values were calculated using the Goal Seek function in Excel. 2.6.2. Sorption isotherms The sorption isotherms were evaluated by different models: Linear, Langmuir, and Freundlich isotherm models. Amount of sorbed antibiotics in soils at equilibrium (q e ) were plotted in axis y versus antibiotics remaining concentration at the equilibrium in the liquid phase (axis x). Excel was used to fit all models, and the correlation coefficients (R 2 ) were applied to evaluate the good fitting of different models to data. Eq. 3 was used to calculate the amount of sorbed pollutants: q e =(C i -C e )⋅V/m (3) where C i (mg L −1 ) and C e (mg L −1 ) are the concentrations of the compound before and after sorption assay, respectively, m is the soil mass (g) and V (L) is the final solution volume. A summary of sorption isotherms models used is in Table 4. 2.6.3. Environmental risk assessment Environmental risk assessment was carried out following the EU technical guideline document [17] by the comparison between measured concentrations and the predicted no-effect concentrations in soil (PNEC soil ). Since toxicity studies on terrestrial organisms (mainly earthworms and plants) are scarce, the equilibrium partitioning method Table 2 Characteristics of soils. Property, units Soil 1 Soil 2 Soil 3 Fine sand (particle sizes from 0.2 to 0.02 mm), wt % 16.5 4.80 14.0 Coarse sand (particle sizes from 2 to 0.2 mm), wt % 8.20 69.5 51.6 Silt (particle sizes from 0.02 to 0.002 mm), wt % 44.5 5.80 18.4 Clay (particle sizes less than 0.002 mm), wt % 30.8 19.9 16.0 pH 8.27 8.21 8.06 EC, µS cm −1 129 75 126 OM, wt % 0.91 0.58 2.01 CEC, meq kg −1 234 200 144 Exchangeable Ca, meq kg −1 140 154 106 Exchangeable Mg, meq kg −1 26.3 9.5 19.5 Exchangeable K, meq kg −1 7.8 2.1 4.6 Exchangeable Na, meq kg −1 <0.5 <0.5 <0.5 CEC: cationic exchange capacity; EC: electrical conductivity; Wt: weight; OM: organic matter. C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 5 can be applied to aquatic data to identify predicted no-effect concentrations (PNEC) for soil organisms. PNEC soil (mg kg −1 ) was calculated from the predicted no-effect concentrations in the aquatic environment (PNEC water ) value (mg L −1 ) multiplied by soil-water distribution coefficient (K d ) (L kg −1 ) of each selected antibiotic. PNEC water was calculated from chronic or acute ecotoxicological data (half maximal effective concentration (EC 50 ), median lethal dose (LC 50 ) or no observed effect concentration (NOEC)) reported in literature for aquatic organisms exposed to selected antibiotics according to EU technical guideline document [45]. Ecotoxicological data can be seen in Table S3. Obtained PNEC soil values are in Table S4. 3. Results and discussion 3.1. Sorption isotherms To investigate the sorption isotherms of antibiotics and their main metabolites on the different soils, Linear, Langmuir, and Freundlich isotherm models were applied to fit the equilibrium data. The Langmuir isotherm model assumes that the surface of the sorbent is covered by many active sites, where each site can be occupied by a sorbed molecule. The sites are all equivalent, and it is believed that the sorbed molecules do not interact with each other or jump from one site to another. Furthermore, sorption is complete when all the sites are occupied, corresponding to a monolayer of sorbate. However, because most surfaces are heterogeneous, there are multiple sites available for sorption, i.e., the sorption varies from one site to another, as described by the Freundlich isotherm model. The Linear isotherm model assumes a homogeneous surface with an infinite number of active sites is assumed, where sorbed molecules do not interact with each other. Thus, sorption is directly proportional to the concentration of sorbate, and it is only applied at low concentrations. This leads to variation in sorption across different sites, resulting in a physical sorption in multilayers. This model accounts for the heterogeneous nature of the sorption process [46]. Parameters derived from these models are summarized in Table 5. Fig. 1 Table 3 Dissipation kinetics models applied. Dissipation kinetics Model Equation Description Single first-order kinetic model (SFO) C =C 0 e −k1t ; t: time (days); k 1 : first-order rate constant (day −1 ); C 0 : initial concentration (µg kg −1 ); C: concentration at time t (µg kg −1 ). This model assumes that the number of molecules in the soil is small in relation to the number of active sites in the soil. In this case, the dissipation of a compound is proportional to the remaining concentration. Bi-exponential model (BEM) C =(k 1 C 0 )/((k 1 +k 2 C 0 )e k1t −k2C0 ); k 1 : first-order rate constant (day −1 ); k 2 : second-order rate constant (kg µg −1 day −1 ); t: time (days); C 0 and C are the concentrations (µg kg −1 ) at the beginning of the experiment and at time t, respectively. This model considers that the degradation occurs in two compartments. The first compartment corresponds to the soil-water phase, where the contaminants are available for microorganisms. In this compartment, a fast dissipation is expected. The second compartment corresponds to a phase where the contaminants are adsorbed to soil particles and the dissipation is controlled by sorption-desorption processes. In this model, the dissipation rate can be represented by the sum of first- and second-order differential equations. The equation is reduced to a single first-order equation when k 2 =0. First-order doubleexponential decay model (FODED) C =C sol e −k1t +C sorb e −k2t ; t: time (day); C sol and C sorb : concentrations of the compounds initially distributed into the solution phase and adsorbed onto soil particles, respectively (µg kg −1 ); k 1 (day −1 ) and k 2 (day −1 ): dissipation rate constants for each compartment. This model considers the dissipation of the compound into two compartments, a solution phase, and a sorbed phase, where the compounds are adsorbed onto soil particles. However, FODED model considers first-order kinetics for both processes. Usually, k 1 is higher than k 2 , indicating a higher degradation rate of the compound in the solution compartment than adsorbed onto soil. Moreover, when k 1 is equal to k 2 , or one of the concentrations (C sol or C sorb ) is zero, the model is mathematically transformed into SFO model. First-order twocompartment model (FOTC) C =C 0 e −(k1 +kr)t +C 0 (k r / (k 1 +k r −k 2 ))(e −k2t −e −(k1 +kr)t ); t: time (day); k r (day −1 ): transfer rate constant of the compound between both compartments; k 1 (day −1 ) and k 2 (day −1 ): rate constants for the fast and slow degradation compartments, respectively; C and C 0 are the concentrations at time t and at t =0 (µg kg −1 ), respectively. This model considers the dissipation of the compounds in a complex system as a heterogeneous soil. This dissipation takes place in a two-compartment system. In one compartment, fast degradation occurs because of physical, photochemical and other processes. In the other compartment, the degradation is slow and mainly due to microbial, chemical or enzymatic processes. FOTC model includes a component of transference between the two compartments not considered in BEM and FODED models. Table 3 (continued) Dissipation kinetics Model Equation Description Logistic model (LM) C =C 0 (a max /(a max −a 0 + a 0 e rt )) amax/r ; a 0 : kinetic constant at the beginning of the batch experiment; a max : maximum value of the kinetic constant; r: microbial growth rate; t: time (day); C and C 0 are the concentrations at time t and at t =0 (µg kg −1 ), respectively. This model considers that the degradation rate constant increases after the application of sludge or compost to a maximum value. Once the kinetic constant has reached its maximum value, the kinetic approach is first order. For this model, the higher the r or a 0 value, the lower the lag phase of the compound. Table 4 Sorption isotherms models studied. Isotherm Model Equation Linear q =Kd ×Ce; Ce: Equilibrium concentration of pollutants (mg L −1 ); q: Equilibrium sorption capacity (mg g −1 ); Kd: solution-soil distribution coefficient (L g −1 ). Langmuir q =(q max ×K L ×C e )/(1 +(C e ×K L ); C e : Equilibrium concentration of pollutants (mg L −1 ); q: Equilibrium sorption capacity (mg g −1 ); q max : maximum amount sorbed within a monolayer (mg g −1 ); K L : Langmuir dissociation constant (L mg −1 ), which is related to the sorption energy. Freundlich q =K F ×C e 1/n ; C e : Equilibrium concentration of pollutants (mg L −1 ); q: Equilibrium sorption capacity (mg g −1 ); K F : Freundlich constant ((mg g −1 )(mg L −1 ) −1/n ), which is related to the affinity of the sorbent to the sorbate; 1/n: dimensionless parameter, which indicates how sorption varies as a function of the concentration. C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 6 shows the sorption isotherms of antibiotics on different soils. The results showed a consistent and similar good fit across all models, although some differences were observed in some cases among antibiotics classes and soils. Overall, all antibiotics, along with their metabolites, fit well into the linear model (R 2 >0.965), except for TMP and its metabolites in soil 3 (R 2 >0.717). Langmuir and Freundlich models also fit in some cases but in a lower extent (R 2 from 0.227 to 0.993 and from 0.524 to 0.999, respectively). These findings suggest that sorption occurs uniformly on the surface through monolayer sorption by partitioning and is in concordance with previous reported data [7,47]. The 1/n values of the Freundlich model were <1.0 for SMX, SMZ, and SDZ and their metabolites, indicating concentration-dependent sorption of these sulfonamides in all three soils, which was not observed for TMP and its metabolites. Additionally, the K d values ranged from 0.002 to 0.065 L g −1 for parent antibiotics and from 0.008 to 0.254 L g −1 for metabolites. The K d values were higher for all metabolites (AcSMX, AcSMZ, AcSDZ, DM-TMP, and 4-OH-TMP) compared to their parent compounds across the studied soils, indicating a slightly higher sorption capacity for the metabolites than the parent compounds. According to the OECD [37], the key parameters of soil that play the most crucial role in the sorption of organic pollutants are the pH of the soil for ionizable compounds, OM content, clay content, soil texture, and cation exchange capacity. The pH is important because antibiotics have pK a values that determine speciation. The pH of the three different soils is very similar, so in this case, no significant differences are shown. Due to the higher pH of the soil (Table 2) and the low pK a of sulfonamides (Table 1), it is likely that sulfonamides were in the anionic form. Their metabolites, however, due to acetylation of the hydroxyl group possess lower negative charge (hence, have a higher pK a ). Anionic organic molecules generally have a low sorption affinity in negatively charged soils (described by cationic exchange capacity) due to repulsion between negative charges [5], this could explain the slightly higher sorption of sulfonamide metabolites with respect to the parent compound. Only in soil 2, the K d values of DM-TMP and OH-TMP were lower than those of TMP, suggesting a higher sorption capacity of this soil for Table 5 Experimental parameters of sorption isotherms of sulfonamide antibiotics and metabolites onto the studied soils. Model Parameter SMX AcSMX SMX-GL SMZ AcSMZ SDZ AcSDZ TMP DM-TMP OH-TMP Soil 1 Linear K D (L g −1 ) 0.008 0.014 n.a. 0.008 0.014 0.005 0.011 0.014 0.024 0.024 R 2 0.991 0.999 n.a. 0.965 0.998 0.994 0.996 0.996 0.970 0.970 Langmuir q max (mg g −1 ) 0.032 0.118 n.a. 0.04 0.142 0.010 0.148 0.051 0.076 0.076 K L (L mg −1 ) 0.259 0.105 n.a. 0.205 0.087 0.433 0.063 0.461 0.560 0.560 R 2 0.977 0.971 n.a. 0.955 0.965 0.978 0.953 0.993 0.920 0.876 Freundlich K F (mg g −1 )(mg L −1 ) −1/n 0.009 0.016 n.a. 0.011 0.018 0.005 0.015 0.016 0.024 0.024 n1.193 1.180 n.a. 1.359 1.262 1.277 1.362 0.778 0.634 0.634 R 2 0.995 0.995 n.a. 0.980 0.992 0.999 0.997 0.997 0.995 0.979 Soil 2 Linear K D (L g −1 ) 0.031 0.246 n.a. 0.065 0.254 0.024 0.249 0.024 0.010 0.010 R 2 0.994 0.978 n.a. 0.984 0.965 0.998 0.973 0.975 0.991 0.991 Langmuir q max (mg g −1 ) 0.196 0.943 n.a. 0.210 2.167 0.020 0.208 0.068 0.035 0.024 K L (L mg −1 ) 0.123 0.039 n.a. 0.265 0.02 0.839 0.161 0.537 0.618 9.635 R 2 0.911 0.255 n.a. 0.944 0.288 0.864 0.227 0.992 0.984 0.740 Freundlich K F (mg g −1 )(mg L −1 ) −1/n 0.049 101.1 n.a. 0.157 410.3 0.063 2.208 0.024 0.023 0.028 n1.323 3.421 n.a. 1.256 3.661 1.668 2.281 0.790 0.277 0.489 R 2 0.992 0.783 n.a. 0.994 0.873 0.986 0.524 0.990 0.954 0.721 Soil 3 Linear K D (L g −1 ) 0.002 0.013 n.a. 0.007 0.021 0.007 0.013 0.007 0.008 0.008 R 2 0.997 0.997 n.a. 0.995 0.991 0.969 0.997 0.717 0.734 0.734 Langmuir q max (mg g −1 ) 0.303 0.118 n.a. 0.022 0.172 0.025 0.147 0.018 0.018 0.028 K L (L mg −1 ) 0.006 0.095 n.a. 0.347 0.105 0.304 0.076 5.346 6.782 17.07 R 2 0.988 0.962 n.a. 0.984 0.953 0.957 0.951 0.857 0.882 0.928 Freundlich K F (mg g −1 )(mg L −1 ) −1/n 0.002 0.015 n.a. 0.007 0.028 0.010 0.017 0.013 0.015 0.025 n1.182 1.211 n.a. 1.091 1.239 1.377 1.308 0.368 0.364 0.19 R 2 0.997 0.994 n.a. 0.995 0.979 0.978 0.988 0.791 0.810 0.833 n.a.: not available Fig. 1. Sorption isotherms of antibiotics on soils. Lines represent the linear curve that best fits the experimental data according to the linear sorption isotherm model. C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 7 TMP. However, only slightly differences between the sorption of the parent compounds and the metabolites were obtained. Moreover, the q max values ranged from 0.01 to 0.19 mg g −1 for soil 1, 0.02 to 2.17 mg g −1 for soil 2, and from 0.02 to 0.3 mg g −1 for soil 3, indicating that sorption follows the order soil 2 >soil 3 >soil 1. This variation could be attributed to the differing physicochemical characteristics and texture of the soils (Table 2) and the particular properties of the antibiotics. Mejías et al. [7] obtained positive correlations between the OM and the sorption of TMP metabolites. Parameter values for the different isotherm models for SMX-GL could not be calculated since the sorption was so high in the soils that the equilibrium concentrations were below the method’s limit of quantification in all cases. In general, TMP and its metabolites were better sorbed to the soils than sulfonamides (K d ranged from 0.002 to 0.021 L g −1 and from 0.007 to 0.024 L g −1 for sulfonamides and TMP and its metabolites, respectively) without considering soil 2, where sulfonamides were better sorbed to the soil than TMP. Zhang et al. [47] also concluded that sulfonamides were found to be weaker sorbed onto all soil samples than TMP, which indicates a higher mobility of sulfonamides due to their different molecular structures. The K d values they obtained ranged from 0.00003 to 0.00047 L g −1 for sulfonamides and from 0.00673 to 0.00921 L g −1 for TMP. Similarly, Koˇ c´ arek et al. [5] also reported a higher sorption for TMP than for SMX. Other studies [48,49] reported higher sulfonamide K d values ranging from 0.00062 to 0.10 L g −1 . 3.2. Content of antibiotics and their metabolites in spiked soils According to literature, TMP has been measured in agricultural soils up to 13.11 μ g kg −1 [50], but no data was reported for its metabolites 4-OH-TMP and DM-TMP. On the other hand, sulfonamides were quantified in higher concentrations, SMX was found at concentrations up to 4460 μ g kg −1 [50] and its metabolite AcSMX in the range from 10.92 to 23.29 μ g kg −1 [51]. However, no data is available for SMX-GL. SDZ was measured in concentrations up to 5560 μ g kg −1 [50] and its metabolite AcSDZ was not detected [52]. SMZ was detected up to 3.5 μ g kg −1 [53]. AcSMZ was not detected [52]. In this study, none of the selected antibiotics and metabolites were found in the studied soils, concentrations were under the detection limit of the method for all compounds in all soils (Table S5). Therefore, batch experiments were carried out spiking the soils with antibiotics and their metabolites at a concentration of 500 μ g kg −1 dm. The concentrations measured after the spiking procedure are summarized in Fig. 2. The pollutants sorption on soil depends on the physical-chemical characteristics of the soils as well as on the physicochemical properties of the compounds. Consequently, differences on sorption onto different soils of some compounds during the spiking procedure can play an important role in the final content of the pollutants in the soils. A low sorption makes the antibiotic available for a rapid degradation during the spiking procedure. This could explain why the contents of the selected compounds, like 4-OH-TMP in spiked soils, measured after the spiking procedure, differed on day 0 of the batch experiment. Mean concentrations were in the ranges 125–473 µg kg −1 in soil 1, 17.2–698 µg kg −1 in soil 2, and 48.6–664 µg kg −1 in soil 3. These concentrations correspond to an efficiency percentage of the spiked concentration from 25.0 to 94.6 % in the case of soil 1, from 3.44 to 140 % for soil 2 and from 9.72 to 132 % in the case of soil 3. Lower concentrations corresponded to sulfonamides metabolites, especially SMX-GL. The differences measured between spiked concentrations and the concentrations measured in spiked soils could be due to the different sorption among target compound onto the soil and the fast degradation of the compounds after the spiked procedure. The differences may also result from interconversion between metabolites and parent compound in soils, leading to an increase in the concentration of some compounds. This phenomenon was observed with SDZ, which could be attributed to the interconversion of AcSDZ to SDZ as it was previously reported [54]. Spiking efficiency percentage were higher in soil 2 for most compounds which could be related to a higher sorption capacity in soil 2 reported previously. However, for SMZ, TMP and SMX-GL the spiking efficiency in soil 2 was the lowest. The sorption of target compounds on the studied soil depends on the physicochemical characteristics of the soil and those of the compounds. For example, the sorption mechanisms include reactions as complexation, or electrostatic interactions, hydrogen bonds or hydrophobic partitions [55,56] and, consequently, they are profoundly influenced by the physicochemical characteristics of the compounds (such as water solubility, functional groups, pK a , Log K ow ). Moreover, several studies have demonstrated the influence of soil physicochemical characteristics such as pH, structure, OM, or clay content on the sorption or mobility of antibiotics in the soil [24,55,57,56] which would explain the differences of the concentrations measured for each compound in the different soils. Furthermore, it is noteworthy that there was no loss of water solution during the homogenization of the soil with the mix spiking solution. Consequently, the low concentrations measured for some of the compounds could be due to their fast degradation during the first hours after the spiking procedure. Moreover, parent compounds were present, for most of the cases, at higher concentrations than their metabolites, except in the case of AcSMZ and DM-TMP and 4-OH-TMP in soil 2. This was particularly relevant in the case of SMX-GL. In general, final spiked concentrations were higher in soil 2, probably due to a higher sorption of compounds onto this soil (Table 5). Higher spiking efficiency is directly correlated to sorption as a high sorption also avoid quickly degradation of compounds as sorbed pollutants are not available Fig. 2. Mean contents of parent compounds and their main metabolites in the fortified soils. C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 8 for degradation. To investigate this, a Partial Least Squares Regression (PLS) analysis was conducted to identify which physicochemical properties of the soils have the greatest impact on antibiotic concentrations and how they are interrelated. The correlation matrix obtained is shown in Fig. S1. From the PLS analysis, it was found that factors such as exchangeable Mg and K, % of Fine Sand, % of Silt, and EC had a negative correlation with the metabolites AcSDZ, AcSMZ, AcSMX, and DM-TMP, suggesting that the concentrations found on the first day are lower with higher values of these factors. This corresponds with soils 1 and 3, as they have a considerably higher % of Fine Sand and % of Silt compared to soil. In contrast, the % of Coarse Sand and exchangeable Ca showed a positive correlation with the concentrations of antibiotics and metabolites, generally aligning with the fact that soil 2 has the highest values. Therefore, the different physicochemical properties could influence the sorption and the concentrations of these metabolites on the first day of the experiment, explaining why they were lower than the spiked concentration and varied between soils. Despite the low concentrations measured for some of the target compounds, these concentrations were enough to assess their dissipation in soils. 3.3. Dissipation of antibiotics and their metabolites in soil 3.3.1. Sulfamethazine and its metabolite Concentrations of SMZ and AcSMZ measured in soils are shown in Figs. S2 to S4. A summary of these concentrations and the applied kinetic models are showed in Fig. 3. The concentrations of SMZ at the beginning of the batch experiment were from 454 (soil 3) to 357 μ g kg −1 dm (soil 2). The fastest dissipation was observed in soil 3, reaching concentrations of about the 10 % of the initial value about 16–21 days from the beginning of the of batch experiments. In soil 2, the concentrations of SMZ remained practically stable for the first 6 weeks (46 days) of the experiment. After this period, the concentrations slightly decreased, ranging from 97 to 259.9 μ g kg −1 dm. The dissipation was more pronounced in the other soils, with concentrations in soil 1 decreasing to between 48.1 and 58.7 μ g kg −1 dm, and concentrations in soil 3 ranging from 8.7 to 11.2 μ g kg −1 dm after 46 days. In the case of the metabolite AcSMZ, a fast dissipation was observed, with concentrations reduced to approximately the 10 % of their initial values after a few days (from 4 to 8 days) from the beginning of the batch experiments in all studied soils. These assessments were corroborated by the study of the kinetic of the process. Optimized parameters of the kinetic models applied to the degradation of the studied compounds are shown in Tables S6 to S8. SMZ showed different dissipation profiles among different soils. Considering soil 1, despite the good fit observed with the SFO kinetic model (R 2 was 0.933 and the slope of the experimental concentration vs model concentrations was 1.01), the experimental data exhibited biphasic behaviour. This means there was a fast dissipation at the beginning of the experiments followed by a slow dissipation. As a result, all biphasic models applied showed a high correlation between experimental concentrations and those obtained from the model (R 2 were 0.942, 0.956, and 0.957 for BEM, FODED, and FOTC, respectively). These results showed the fastest dissipation at the beginning of the experiments (for example, k 1 values in FODED and FOTC model were 0.077 d −1 ), followed by a slow dissipation stage (k 2 were 0.004 d −1 ), resulting in a stable concentration (58.7 µg kg −1 ) at the final stage of the experiment. This could be explained by the distribution of SMZ into two phases: the aqueous solution phase, where rapid decomposition occurs probably due to the large availability for the microorganisms involved in the dissipation process, and the sorbed phase, where the degradation depends on the soil-water partition equilibriums. This behaviour has been previously described in the case of SMZ in water-sediment systems [58] obtaining values of k 1 and k 2 of 0.20 and 0.03 d −1 , respectively, with a R 2 of 0.999. It has been postulated that biphasic dissipation patterns result from reduced microbial activity toward the end of long-term dissipation studies because it is assumed that the target compound must be in the aqueous phase to be available for biodegradation and that no degradation occurs in the sorbed phase (Srinivasan et al., 2014). The first dissipation phase is fast because there is more fraction of compound in the aqueous phase, and therefore, it is bioavailable for degradation. As the compound degrades from the aqueous phase the degradation kinetics will shift to a slower phase as degradation is dependent on prior desorption of the compound from the soil particles into the aqueous phase. Therefore, compounds that are highly sorbed on soil are more likely to have biphasic dissipation kinetics, as there will be a fraction of the compound in the liquid phase and another fraction sorbed to the solid phase. However, compounds that do not have a high affinity to sorb onto soil will be mostly in the liquid phase and, therefore, are likely to have a SFO kinetic model. Therefore, soils in which antibiotics have a higher K d are more susceptible to biphasic kinetics. Fig. 3. Mean concentrations of SMZ and AcSMZ as a function of time in soil 1, soil 2 and soil 3 with curves from kinetic models. C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 9 In the case soil 3, the dissipation on SMZ mainly followed a SFO kinetic (BEM and LM model were transformed to SFO). However, FODED and FOTC models showed a good fit with the experimental data (R 2 values were 0.965 and 0.967 for FODED and FOTC model, respectively). This good fit for the SFO and biphasic model in the case of soil 3 could be explained by its faster dissipation (k 1 obtained in FODED model was 0.011 d −1 ) compared to that observed in soil 1 and soil 2 (where k 1 were 0.076 and 0.279 d −1 , respectively). Consequently, the lower final concentrations achieved in soil 3 allowed a better fit to the SFO model. For instance, while concentrations measured in soil 1 ranged from 48.1 to 57.8 µg kg −1 after day 46, those in soil 3 were notably lower, ranging between 8.67–11.7 µg kg −1 . In the case of soil 2, the dissipation of SMZ followed a LM kinetic. The concentrations remain approximately constant (from 366 to 359 µg kg −1 ) for the first 25 days followed by a SFO kinetic dissipation (Fig. 3). This behaviour could be explained by considering two hypotheses: first, a stronger sorption of SMZ on soil 2 resulting in a lower availability for soil microorganism and consequently, slower dissipation at the beginning of the batch experiment; second, the dissipation of SMZ in soil 2 could be preceded by a lag phase due to a microbial adaptation to SMZ. This fact has been previously described in the microbial degradation of structurally related sulfonamides in sewage sludge [59-61]. The estimated endpoint values for each model can be found in Tables S6 to S8, while the average values are illustrated in Fig. S5 of the supplementary material. DT 50 values can also be seen in Table 6. Endpoint values of SMZ followed the order soil 3 <soil 1 <soil 2. Considering the models that showed the best fit to the experimental data, the following endpoints were estimated: DT 50 values decreased from soil 2 (66.9 days for LM), to soil 1 (11.1 and 11.04 days for FODED and FOTC), to soil 3 (2.75 and 3.54 days for FODED and FOTC); DT 90 values decreased from soil 2 (147 days), to soil 1 (77.8 and 82.5 days), and to soil 3 (11.0 and 13.3 days) for these models. These DT 50 values were in concordance with those previously reported in other soils as silt loam and sandy soils [59,62] and clay loam soil [62-64], obtaining DT 50 values from 1.2 to 77 days (Table 6). Besides, the biphasic kinetic of the dissipation of SMZ has been described previously on water-sediment system [58] In this work, the authors described the appearance of a fast degradation in the first days (k 1 of 0.20 d −1 ) follow by a second phase on the dissipation kinetic (k 2 0.03 d −1 ) as result of the addition of manure, mainly due to an increase of the OM. Similar results were obtained by Lertpaitoonpan et al. [65] in soil amended with manure, in their study SMZ disappeared rapidly during the first 7 days followed by slow disappearance which may indicate that SMZ had become more persistent and less available. However, although degradation of sulfonamides including SMZ preceded by a lag phase has been previously reported in sewage sludge [60,61], to the best of our knowledge, there are no studies in recent scientific literature that report this kinetic dissipation for SMZ in soil. Regarding to AcSMZ, this compound showed a fast dissipation in all studied soils being almost fully degraded in only 7 days. It followed a Table 6 Obtained DT 50 values (days) and literature values. SDZ AcSDZ SMZ AcSMZ SMX AcSMX SMXGL TMP 4-OH- TMP DMTMP Reference Soil 1 SFO 6.19 3.54 14.4 0.75 4.41 4.01 4.36 49.5 11.6 15.4 This study BEM 2.24 3.98 12.7 0.34 4.72 2.94 4.15 49.5 7.43 11.6 This study FODED 4.15 0.96 11.1 0.75 4.41 1.45 3.35 49.5 7.27 12.4 This study FOTC 4.36 2.17 11.0 0.11 3.91 2.04 3.34 47.9 7.17 11.9 This study LM 7.53 5.74 16.1 0.77 6.03 3.28 6.30 53.9 8.25 13.9 This study Soil 2 SFO 34.7 3.41 63.0 2.11 40.8 3.27 25.7 >120 31.5 4.88 This study BEM 40.8 3.98 38.2 2.01 49.5 4.41 17.8 >120 33.0 4.88 This study FODED 34.7 2.49 63.0 1.80 40.8 2.97 12.7 >120 31.5 4.88 This study FOTC 42.8 4.80 36.7 1.79 44.6 2.88 11.3 >120 34.1 4.64 This study LM 41.7 4.63 66.9 3.14 45.8 4.85 18.2 >120 31.7 7.90 This study Soil 3 SFO 2.84 1.51 3.32 1.10 5.64 10.1 2.97 14.1 3.40 3.50 This study BEM 2.54 3.87 3.32 0.89 3.61 3.46 3.04 14.2 5.02 4.28 This study FODED 2.67 1.11 2.75 0.91 3.13 0.93 1.29 14.1 3.13 1.24 This study FOTC 2.67 1.51 3.54 0.09 3.13 0.93 0.94 18.0 3.13 3.27 This study LM 5.33 2.62 3.32 1.10 6.13 29.1 2.53 13.3 2.87 2.87 This study Literature values Matrix Soils 18.6 Accinelli et al.[59] Manure-amended soil 1.2 −63 Lertpaitoonpan et al. [65] Soils 1.3 −63 Lertpaitoonpan et al. [65] Soils 24.8 Pan and Chu,[63] Soils 17.8 −30.1 Moody et al.,[62] Manure-amended soil 8.48 −10.22 10.28 −13.68 Zhang et al.[66] Soils 0.3 −13.3 Kodeˇ sov´ a et al.[67] Soils 20.83 −38.80 28.85 −36.04 Shen et al.[68] Soils 53 77 1 Sentek et al.,[69] Manure digestateamended soil 7 Rauseo et al.,[70] Soils 4.3 −25.4 Srinivasan and Sarmah,[18] Soils 3.98 −6.13 1.20 −5.62 Santos et al.[4] Soils 21.1 −23.1 1.65 −1.99 Malvar et al.[26] Soils 15.3 −821 Kodeˇ sov´ a et al.[71] Soils 12 −75 Berendsen et al.[72] Soils 62 −161 Wu et al.[73] Soils 19 −790 Rietra et al.[74] C. Mejías et al. Journal of Hazardous Materials 479 (2024) 135586 16 2, 2024. Available from: 〈https://echa.europa.eu/documents/10162/987906/t gdpart2_2ed_en.pdf/138b7b71-a069–428e-9036–62f4300b752f〉. [46] Wang, J., Guo, X., 2020. Adsorption isotherm models: Classification, physical meaning, application and solving method. Chemosphere 258, 127279. https://doi. org/10.1016/j.chemosphere.2020.127279. [47] Zhang, Y.L., Lin, S.S., Dai, C.M., Shi, L., Zhou, Z.F., 2014. Sorption-desorption and transport of trimethoprim and sulfonamide antibiotics in agricultural soil: effect of soil type, dissolved organic matter, and pH. Environ Sci Pollut Res 21, 5827–5835. https://doi.org/10.1007/s11356-014-2493-8. [48] Díaz-Cruz, M.S., L´ opez de Alda, M.J., Barcel´ o, D., 2003. Environmental behavior and analysis of veterinary and human drugs in soils, sediments and sludge. Trends Anal Chem 22, 340–351. https://doi.org/10.1016/S0165-9936(03)00603-4. [49] Thiele-Bruhn, S., Seibicke, T., Schulten, H.R., Leinweber, P., 2004. Sorption of sulfonamide pharmaceutical antibiotics on whole soils and particle-size fractions. J Environ Qual 33 (4), 1331–1342. https://doi.org/10.2134/jeq2004.1331. [50] Fang, L., Chen, C., Li, S., Ye, P., Shi, Y., Sharma, G., Sarkar, B., Shaheen, S.M., Lee, S.S., Xiao, R., Chen, X., 2023. A comprehensive and global evaluation of residual antibiotics in agricultural soils: Accumulation, potential ecological risks, and attenuation strategies. Ecotoxicol Environ Saf 262, 115175. https://doi.org/ 10.1016/j.ecoenv.2023.115175. [51] Liu, X., Liang, C., Liu, X., Zhao, F., Han, C., 2020. Occurrence and human health risk assessment of pharmaceuticals and personal care products in real agricultural systems with long-term reclaimed wastewater irrigation in Beijing, China. Ecotoxicol Environ Saf 190, 110022. https://doi.org/10.1016/j. ecoenv.2019.110022. [52] García-Gal´ an, M.J., Díaz-Cruz, S., Barcel´ o, D., 2013. Multiresidue trace analysis of sulfonamide antibiotics and their metabolites in soils and sewage sludge by pressurized liquid extraction followed by liquid chromatography–electrosprayquadrupole linear ion trap mass spectrometry. J Chromatog A 1275, 32–40. https://doi.org/10.1016/j.chroma.2012.12.004. [53] Stando, K., Wilk, J., Jak´ obik-Kolon, A., Felis, E., Bajkacz, S., 2023. Application of UHPLC-MS/MS method to monitor the occurrence of sulfonamides and their transformation products in soil in Silesia, Poland. Environ Sci Pollut Res 30, 112922–112942. https://doi.org/10.1007/s11356-023-30146-y. [54] Bioˇ si´ c, M., Mitrevski, M., Babi´ c, S., 2017. Environmental behavior of sulfadiazine, sulfamethazine, and their metabolites. Environ Sci Pollut Res 24, 9802–9812. https://doi.org/10.1007/s11356-017-8639-8. [55] Cela-Dablanca, R., Barreiro-Buj´ an, A., Ferreira-Coelho, G., Rodríguez-L´ opez, L., Sant´ as-Miguel, V., Arias-Est´ evez, M., et al., 2022. Competitive adsorption and desorption of tetracycline and sulfadiazine in crop soils. Environ Res 214, 113726. https://doi.org/10.1016/j.envres.2022.113726. [56] Lin, Q., Li, B., Liu, X., Zhang, B., Xu, S., 2022. Insights into sorption and leaching behaviour of sulfadiazine in soil as affected by humic acid. J Soil Sediment 22, 809–817. https://doi.org/10.1007/s11368-021-03110-x. [57] Fan, Z., Casey, F.X.M., Hakk, H., Larsen, G.L., Khan, E., 2011. Sorption, fate, and mobility of sulfonamides in soils. Water Air Soil Pollut 218, 49–61. https://doi. org/10.1007/s11270-010-0623-6. [58] Carstens, K.L., Gross, A.D., Moorman, T.B., Coats, J.R., 2013. Sorption and photodegradation processes govern distribution and fate of sulfamethazine in freshwater – sediment microcosms. Environ Sci Technol 47, 10877–10883. https:// doi.org/10.1021/es402100g. [59] Accinelli, C., Koskinen, W.C., Becker, J.M., Sadowsky, M.J., 2007. Environmental fate of two sulfonamide antimicrobial agents in soil. J Agric Food Chem 55, 2677–2682. https://doi.org/10.1021/jf063709j. [60] Ingerslev, F., Halling-Sørensen, B., 2000. Biodegradability properties of sulfonamides in activated sludge. Environ Toxicol Chem 19, 2467–2473. https:// doi.org/10.1002/etc.5620191011. [61] P´ erez, S., Eichhorn, Aga, D.S., 2009. Evaluating the biodegradability of sulfamethazine, sulfamethoxazole, sulfathiazole, and trimethoprim at different stages of sewage treatment. Environ Toxicol Chem 24 (6), 1361–1367. https://doi. org/10.1897/04-211R.1. [62] Moody, A.H., Lerch, R.N., Goyne, K.W., Anderson, S.H., Mendoza-C´ ozatl, D.G., Alvarez, D.A., 2024. Degradation kinetics of veterinary antibiotics and estrogenic hormones in a claypan soil. Chemosphere 346, 140501. https://doi.org/10.1016/j. chemosphere.2023.140501. [63] Pan, M., Chu, L.M., 2015. Adsorption and degradation of five selected antibiotics in agricultural soil. Sci Total Environ 545-546, 48–56. https://doi.org/10.1016/j. scitotenv.2015.12.040. [64] Pan, M., Chu, L.M., 2017. Leaching behaviour of veterinary antibiotics in animal manure-applied soils. Sci Total Environ 579, 466–473. https://doi.org/10.1016/j. scitotenv.2016.11.072. [65] Lertpaitoonpan, W., Moorman, T.B., Ong, S.K., 2015. Effect of swine manure on sulfamethazine degradation in aerobic and anaerobic soils. Water Air Soil Pollut 226, 81. https://doi.org/10.1007/s11270-014-2286-1. [66] Zhang, Y., Hu, S., Zhang, H., Shen, G., Yuan, Z., Zhang, W., 2017. Degradation kinetics and mechanism of sulfadiazine and sulfamethoxazole in an agricultural soil system with manure application. Sci Total Environ 607–608, 1348–1356. https://doi.org/10.1016/j.scitotenv.2017.07.083. [67] Kodeˇ sov´ a, R., Chroˇ n´ akov´ a, A., Grabicov´ a, K., Koˇ c´ arek, M., Schmidtov´ a, Z., Frkov´ a, Z., et al., 2020. How microbial community composition, sorption and simultaneous application of six pharmaceuticals affect their dissipation in soils. Sci Total Environ 746, 141134. https://doi.org/10.1016/j.scitotenv.2020.141134. [68] Shen, G., Zhang, Y., Hu, S., Zhang, H., Yuan, Z., Zhang, W., 2018. Adsorption and degradation of sulfadiazine and sulfamethoxazole in an agricultural soil system under an anaerobic condition: Kinetics and environmental risks. Chemosphere 194, 266–274. https://doi.org/10.1016/j.chemosphere.2017.11.175. [69] Sentek, V., Braun, G., Braun, M., Sebesvari, Z., Renaud, F.G., Herbst, M., et al., 2020. Salinity-independent dissipation of antibiotics from flooded tropical soil: a microcosm study. Sci Rep 10, 14088. https://doi.org/10.1038/s41598-020-70943- w. [70] Rauseo, Barra Caracciolo, J., Ademollo, A., Cardoni, N., Lenola, M., Di, M., Gaze, W., et al., 2019. Dissipation of the antibiotic sulfamethoxazole in a soil amended with anaerobically digested cattle manure. J Hazard Mater 378, 120769. https://doi.org/10.1016/j.jhazmat.2019.120769. [71] Kodeˇ sov´ a, R., Koˇ c´ arek, M., Klement, A., Golovko, O., Koba, O., F´ er, M., et al., 2016. An analysis of the dissipation of pharmaceuticals under thirteen different soil conditions. Sci Total Environ 544, 369–381. https://doi.org/10.1016/j. scitotenv.2015.11.085. [72] Berendsen, B.J.A., Roelofs, G., van Zanten, B., Driessen-van Lankveld, W.D.M., Pikkemaat, M.G., Bongers, I.E.A., et al., 2021. A strategy to determine the fate of active chemical compounds in soil; applied to antimicrobially active substances. Chemosphere 279, 130495. https://doi.org/10.1016/j.chemosphere.2021.130495. [73] Wu, Y., Williams, M., Smith, L., Chen, D., Kookana, R., 2012. Dissipation of sulfamethoxazole and trimethoprim antibiotics from manure-amended soils. J Environ Sci Health B 47, 240–249. https://doi.org/10.1080/ 03601234.2012.636580. [74] Rietra, R.P.J.J., Berendsen, B.J.A., Mi-Gegotek, Y., R¨ omkens, P.F.A.M., Pustjens, A. M., 2024. Prediction of the mobility and persistence of eight antibiotics based on soil characteristics. Heliyon 10, e23718. https://doi.org/10.1016/j.heliyon.2023. e23718. [75] Branco, R.H.R., Meulepas, R.J.W., Kadlecov´ a, K., Cardoso, M.F.S., Rijnaarts, H.H. M., Sutton, N.B., 2024. Effect of dissolved organic carbon on micropollutant biodegradation by aquifer and soil microbial communities. Chemosphere 347, 140644. https://doi.org/10.1016/j.chemosphere.2023.140644. [76] Bartnik, M., Facey, P.C., 2017. Glycosides. Pharmacognosy. In: Badal, S., Delgoda, R. (Eds.), Fundamentals, Applications and Strategies. https://doi.org/ 10.1016/B978-0-12-802104-0.00008-1. [77] Geng, C., Bergheaud, V., Garnier, P., Zhu, Y.G., Haudin, C.S., 2018. Impact of sludge treatments on the extractability and fate of acetyl sulfamethoxazole residues in amended soils. Chemosphere 194, 828–836. https://doi.org/10.1016/j. chemosphere.2017.12.003. [78] Goulas, A., Sertillanges, N., Brimo, K., Garnier, P., Bergheaud, V., Dum´ eny, V., et al., 2019. Environmental availability of sulfamethoxazole and its acetylated metabolite added to soils via sludge compost or bovine manure. Sci Total Environ 651 (Pt 1)), 506–515. https://doi.org/10.1016/j.scitotenv.2018.09.100. [79] Ji, Z., Tang, W., Pei, Y., 2022. Constructed wetland substrates: A review on development, function mechanisms, and application in contaminants removal. Chemosphere 286, 131564. https://doi.org/10.1016/j.chemosphere.2021.131564. [80] Hijosa-Valsero M., Reyes-Contreras C., Domínguez C., B´ ecares E., Bayona J.M. Behaviour of pharmaceuticals and personal care products in constructed wetland compartments: Influent, effluent, pore water, substrate and plant roots. Chemosphere 2016;145:508–517. doi:10.1016/j.chemosphere.2015.11.090. [81] Li, W.C., 2014. Occurrence, sources, and fate of pharmaceuticals in aquatic environment and soil. Environ Pollut 187, 193–201. https://doi.org/10.1016/j. envpol.2014.01.015. C. Mejías et al.